On 2010/10/15 10:42, Stuart Henderson wrote: > $ snmpwalk -v2c -c public $SOMEHOST OPENBSD-MEM-MIB::memIfName > <segfault> > > introduced when per-interface livelock counters were removed. > the variable should probably be removed completely sometime, but is > this ok for now?
it probably would help if I include the backtrace too. #0 0x07bbfb51 in strlen () from /usr/lib/libc.so.57.0 #1 0x1c0050ff in ber_add_string (prev=0x7cce1260, string=0x0) at /usr/src/usr.sbin/snmpd/ber.c:263 #2 0x1c00deb0 in mib_memiftable (oid=0x3c010508, o=0xcfbb99d0, elm=0xcfbb9970) at /usr/src/usr.sbin/snmpd/mib.c:1431 #3 0x1c009acc in mps_getnextreq (root=0x7cce1240, o=0xcfbb99d0) at /usr/src/usr.sbin/snmpd/mps.c:237 #4 0x1c008a84 in snmpe_parse (ss=0xcfbca284, root=0x7cce10e0, msg=0xcfbca068) at /usr/src/usr.sbin/snmpd/snmpe.c:637 #5 0x1c00904a in snmpe_recvmsg (fd=6, sig=2, arg=0x867bd7b0) at /usr/src/usr.sbin/snmpd/snmpe.c:752 #6 0x019810b2 in event_base_loop (base=0x8382e800, flags=0) at /usr/src/lib/libevent/event.c:402 #7 0x01981369 in event_loop (flags=0) at /usr/src/lib/libevent/event.c:478 #8 0x0198138e in event_dispatch () at /usr/src/lib/libevent/event.c:416 #9 0x1c007fba in snmpe (x_env=0x867bd7b0, pipe_parent2snmpe=0x3c01d9e8) at /usr/src/usr.sbin/snmpd/snmpe.c:167 #10 0x1c0167a2 in main (argc=0, argv=0xcfbca6bc) at /usr/src/usr.sbin/snmpd/snmpd.c:184 > > Index: mib.c > =================================================================== > RCS file: /cvs/src/usr.sbin/snmpd/mib.c,v > retrieving revision 1.42 > diff -u -p -r1.42 mib.c > --- mib.c 23 Sep 2010 10:49:55 -0000 1.42 > +++ mib.c 15 Oct 2010 09:39:39 -0000 > @@ -1428,7 +1428,8 @@ mib_memiftable(struct oid *oid, struct b > ber = ber_add_string(ber, kif->if_name); > break; > case 2: > - ber = ber_add_string(ber, 0); > + ber = ber_add_integer(ber, 0); > + ber_set_header(ber, BER_CLASS_APPLICATION, SNMP_T_COUNTER64); > break; > default: > return (-1);
